Expected R2
The Expected R² (Coefficient of Determination or "R squared") is the evaluation of the error made when using the predictive model to estimate the future values of the target, over the horizon.
The R² is the proportion of the variation of the target variable that is explained by the predictive model.
An Expected R² of 1 indicates a predictive model that perfectly captures the variation of the signal. Note that R² does not indicates a goodness of fit. A model with a high R² can predict the variation of the signal but not necessarily be precise on its level.